Final species and habitat distributions for current and future state

作者：Sara Si‐Moussi, Marianne Tzivanopoulos, Gabrielle Deschamps, Maxime Hoareau, Julien Renaud, Rémi Lemaire‐Patin, Wilfried Thuiller · 年份：2025 · DOI：10.3897/arphapreprints.e180864 · 被引用次数：1 · 研究领域：Species Distribution and Climate Change、Environmental DNA in Biodiversity Studies、Wildlife-Road Interactions and Conservation

This deliverable outlines the creation of high-resolution (1km²) distribution maps for species and habitats across Europe, crucial for biodiversity conservation, policy compliance, and ecosystem management. Employing advanced Species Distribution Models (SDMs) and Habitat Distribution Models (HDMs), the task addressed plants, vertebrates, invertebrates, and all EUNIS Level 3 habitats. Species distribution modeling involved machine learning algorithms, carefully selected environmental variables, and spatially comprehensive occurrence datasets from GBIF, EVA, and other databases. Ensemble modeling techniques, spatial block cross-validation, and pseudo-absence generation ensured robust, reliable predictions, validated with metrics like True Skill Statistic (TSS). Habitat modeling similarly utilized environmental predictors (climate, topography, hydrography, geology, soil properties) alongside vegetation plot data from EVA and additional regional databases. Multi-class classification and ensemble forecasting methods provided high-quality predictive habitat maps validated externally and through cross-validation. Current and future scenarios (2050) were developed under varying climate and land-use trajectories (SSP1-RCP2.6, SSP3-RCP7.0), incorporating model uncertainty and expert-informed constraints. These maps support targeted conservation planning, monitoring programs, and decision-making, guiding efforts to enhance Europe's protected area network and biodiversity management.